Can ChatGPT read your website? Check free →
Check my site
How we differ

How we differ: nine facts you can check

Each is a fact about how we work, with the date it was last checked and the place you can check it yourself.

  1. Every number we print carries its source, the date it was read and how it was measured.
    As of 2026-11-03. Check it: The complete sample audit
  2. We invent no authority score of our own. Where we show another company's metric, it is named as theirs, with the date we read it.
    As of 2026-11-03. Check it: How we score
  3. Our reports use no rating words. A number is shown with the arithmetic behind it, and the formula is published.
    As of 2026-11-03. Check it: The OnRank Score formula
  4. Every rule a finding rests on comes from the primary documentation — Google's, Bing's or the platform's own — and is either quoted word for word or named and linked without quotation marks.
    As of 2026-11-03. Check it: The evidence pages in the sample
  5. A person checks and signs every report before it is sent.
    As of 2026-11-03. Check it: What you get in an audit
  6. We run 121 named checks on every site we crawl, and show every issue they found, counted as distinct pages or addresses, with the pages each is found on. Nothing is held back to make a report look better or worse.
    As of 2026-11-11. Check it: “Every check we ran” in the sample
  7. The free check is instant and wall-free: no email, no account, never a password.
    As of 2026-11-03. Check it: The free first look · The free tools
  8. We publish our prices, and print each beside what it pays for.
    As of 2026-11-03. Check it: What you receive, and what it costs · Done for you
  9. We report AI visibility as a mention rate over repeated runs of the same questions, with the run count beside it — never one answer.

    Two published studies are why. University of St. Gallen researchers, arXiv 2604.07585 (April 2026) report a standard error of 0.370 for a single run and recommend at least seven runs per prompt per day. SparkToro with Gumshoe.ai (January 2026) had 600 volunteers run 12 prompts a combined 2,961 times and reports a less than 1 in 100 chance that ChatGPT or Google's AI returns the same list of brands in any two responses.